We now have preliminary data showing that HLA-A alleles exhibit allele-specific variation in mRNA expression levels. Unfortunately, no specific antibodies are available to fully confirm that this diversity is reflected at the protein level. However, we have identified two monoclonal antibodies that allowed us to show a correlation between mRNA and cell surface protein expression levels in a subset of individuals with specific HLA-A genotypes. Several studies have identified epigenetic mechanisms, specifically DNA methylation, as a contributing factor for varied expression levels within the MHC region, but DNA methylation patterns at the allelic level have not been examined. We hypothesized that DNA methylation may explain, at least in part, the differential expression levels of HLA-A mRNA. We have now shown differential DNA methylation across HLA-A alleles, which may explain allele-specific mRNA expression level variation. Higher methylation levels were observed for HLA-A alleles with low expression levels, consistent with the notion that higher methylation levels inhibit transcription and subsequently results in lower gene expression levels. Thus far, our studies have been limited to individuals of European descent, a population with distinct HLA alleles/allele frequencies compared to those of African descent. In collaboration with Drs. Thumbi Ndungu and Salim Karim at the University of Kwa Zulu Natal, we have now acquired a cohort of 280 black South Africans and plan to determine allelic expression levels in this population. This will allow us to assign expression coefficients to disease cohorts composed of African background. We also have data suggesting that there has been selection pressure to maintain differential expression levels of HLA-C and potentially HLA-A. Interestingly, our preliminary data show that HLA-A mRNA expression levels correlate positively with those of HLA-C, but only among subjects carrying miR-148a inhibited HLA-C alleles (I/I) and not among those that escape miR-148a inhibition (D/D). We have also analyzed genotype data in 27 worldwide populations from several ethnic backgrounds using expression level data derived from Caucasians and we observed a similar correlation. We have also begun to analyze HLA-B expression levels. Unlike HLA-C and HLA-A, no difference in mRNA expression levels was seen across HLA-B alleles. There is no single antibody available that recognizes all HLA-B allotypes with equal affinity, but does not recognize HLA-A or -C allotypes. We have used two antibodies, each of which was able to detect two or more HLA-B alleles with equal affinity as evaluated using LABScreen magnetic beads. Based on our screening thus far, there appears to be very little variation in cell surface expression of HLA-B, but our efforts are ongoing. Differential expression levels for allotypes of the HLA class I or II loci may explain some of the many strong associations observed between variation at the HLA loci and disease outcomes. Precise knowledge of the relative protein expression levels of the four HLA class I loci expressed on most cells, HLA-A, HLA-B, HLA-C and HLA-E, seems fundamental in this regard. This is also of interest on HIV-infected cells, as it is not known which HLA locus dominates on HIV infected cells, thereby potentially mediating the greatest protective T cell responses. Mass spectrometry has the ability to discriminate highly homologous proteins and is able to quantitate each locus with equal efficiency, in contrast to antibody-mediated approaches. Thus far we have demonstrated that the procedure can work using an EBV-transformed B cell line. Precipitated HLA was digested with trypsin and the peptide fragments produced were ionized and analyzed by mass spectrometry. Multiple peptide fragments unique to each of the four HLA proteins present in these individuals were detected and for each of 16 peptides (3-5 peptides specific to each locus), their abundance was quantified. Thus, we have shown that we can discriminate and quantify by spectrometry all 4 HLA molecules present in these donors. We now aim to perform the assay using HLA precipitated from normal PBMCs and from in vitro HIV infections of primary CD4 cells.